Yep, JS is correct, they should have used two low jitter clocks to generate the audio clocks. As far as I can tell their business strategy is to release a product later at a higher price that is further up on the audio ladder - I've gleaned this by reading their forum replies translated by Google

. They consider the existing product range just for mass market appeal.
So the question is - how bad is the frequency synthesiser used? Normally, these add a lot of jitter to the clock. In this case, it seems not to be so bad because it's using the Xilinx DCM (digital clock management) in the FPGA. JS did a good analysis of this also here:
DiyHiFi graph of Musiland Vs AP SYS2722
His conclusion of 300ps jitter at the output pins is probably considered good, but not excellent (he doesn't consider it good). This is compared to the 970ps being reported for the M2Tech at the output. Edit: slim.a pointed out that this was an unfaircomparison as the M2tech jitter was a measured value at the output of the SPDIF whereas the Musiland was a theoretical jitter value at the FPGA pin output
This 300ps is also the max value that it could be with all gates in the FPGA firing, etc. Another view on this comes from some measurements taken
http://go.head-fi.org/?id=475X763&ur...40486%23p40486 where the noise of the Musiland was plotted against the noise from an AP SYS2722 and showed it only slightly higher (2-5dB). Interestingly, the SYS2722 was(is?) the equipment used by Stereophile to do it's jitter measurements on equipment. Other measurements were done by FMAK on Audioasylum which indicated a jitter measure of about 100ps.
It would be good to measure the Musiland & M2Tech & whatever else with the same equipment to get a relative ranking of jitter from these units as using absolute jitter figures is fraught with danger.
Anyway, I'm changing the crystal in the Musiland for a low jitter clock (with good PS) & will see if it makes a difference to the sound
